preparing a new release (part 2)

preparing a new release (part 2)

[Brian Gough]

I've made second version to fix some of the errors reported (memory leak in
tests and jacobi svd) You can test it here:

http://www.network-theory.co.uk/download/gsl/gsl-1.6.92.tar.gz
http://www.network-theory.co.uk/download/gsl/gsl-1.6.92.tar.gz.sig

Thanks,

-- 
Brian Gough

Re: preparing a new release (part 2)

[Kevin Bube]

[-- Attachment #1: Type: text/plain, Size: 337 bytes --]

Hello,

'./configure && make && make check' works on AMD64 with gcc 3.3.3.
On dual alpha with gcc 2.96 it works with CFLAGS=-O1.

No change to the previous release here.


Regards,

Kevin

-- 
publickey 2048R/0AFDFB19: http://www.icbm.de/~bube/publickey.asc
fingerprint: 542B 1378 04AA AF1F 572E  78BF 1BF5 5C71 0AFD FB19

[-- Attachment #2: Type: application/pgp-signature, Size: 480 bytes --]

Re: preparing a new release (part 2)

[Brian Gough]

Thanks for the testing -- gsl-1.7 is now released.

Re: preparing a new release (part 2)

[James Theiler]

make and make check worked (again) on RHEL 3
and i noticed that the documentation errors
have been fixed, and that binomial has been
updated.  thanks (again!) for all your efforts.

jt



On Tue, 6 Sep 2005, Brian Gough wrote:

] I've made second version to fix some of the errors reported (memory leak in
] tests and jacobi svd) You can test it here:
] 
] http://www.network-theory.co.uk/download/gsl/gsl-1.6.92.tar.gz
] http://www.network-theory.co.uk/download/gsl/gsl-1.6.92.tar.gz.sig
] 
] Thanks,
] 
] 

-- 
James Theiler            Space and Remote Sensing Sciences
MS-B244, ISR-2, LANL     Los Alamos National Laboratory
Los Alamos, NM 87545     http://nis-www.lanl.gov/~jt

Re: preparing a new release (part 2)

[Jonathan G. Underwood]

Brian Gough wrote:
> I've made second version to fix some of the errors reported (memory leak in
> tests and jacobi svd) You can test it here:
> 
> http://www.network-theory.co.uk/download/gsl/gsl-1.6.92.tar.gz
> http://www.network-theory.co.uk/download/gsl/gsl-1.6.92.tar.gz.sig
> 
> Thanks,
> 

succesful build and check on a Fedora Core 4 system:
gcc version 4.0.1 20050727 (Red Hat 4.0.1-5)
glibc-2.3.5

Checked with CFLAGS=-O2 and CFLAGS=-O3. No problems.

Jonathan.

Re: preparing a new release (part 2)

[Andris Pavenis]

On Tuesday 06 September 2005 14:22, Brian Gough wrote:
> I've made second version to fix some of the errors reported (memory leak in
> tests and jacobi svd) You can test it here:
>
> http://www.network-theory.co.uk/download/gsl/gsl-1.6.92.tar.gz
> http://www.network-theory.co.uk/download/gsl/gsl-1.6.92.tar.gz.sig
>
> Thanks,

Builds OK for DJGPP (v2.04 beta, 2,03 not tested, gcc-4.0.1, WinXP Pro+SP2).
Also all tests passes if the following small EXEEXT related patch is applied 
(it was for 1.6, but it  applies also to 1.6.92):

Andris


--- gsl-1.6/wavelet/Makefile.am~1	2004-07-23 16:55:12.000000000 +0000
+++ gsl-1.6/wavelet/Makefile.am	2005-04-29 13:11:56.000000000 +0000
@@ -6,10 +6,10 @@
 
 libgslwavelet_la_SOURCES = dwt.c wavelet.c bspline.c daubechies.c haar.c
 
-TESTS = test
-
 check_PROGRAMS = test
 
+TESTS = $(check_PROGRAMS)
+
 test_LDADD = libgslwavelet.la ../blas/libgslblas.la ../cblas/libgslcblas.la ../matrix/libgslmatrix.la ../vector/libgslvector.la ../block/libgslblock.la ../ieee-utils/libgslieeeutils.la  ../err/libgslerr.la ../test/libgsltest.la ../sys/libgslsys.la ../utils/libutils.la
 
 test_SOURCES = test.c
-

Re: preparing a new release (part 2)

[linas]

On Wed, Sep 07, 2005 at 06:14:41PM +0300, Andris Pavenis was heard to remark:
>  
> -TESTS = test
> -
>  check_PROGRAMS = test
>  
>  test_SOURCES = test.c

This is a really bad name for a unix prgram. For, on my machine

  which test 

returns

  /usr/bin/test

-------------
I once had a second line manager who wanted to show off his programming
abilities. He wrote a demo program, and he went to show it off. And,
in front of 50 people, it didn't work. We all craned over the monitor
trying to figure out what was wrong (for it looked spotless).  Until
some wizened soul in the back of the crowd cleared his throat and 
asked: "What did you NAME your program?" "I named it 'test'. Why
do you want to know?"

--linas

Re: preparing a new release (part 2)

[Philip Kendall]

On Wed, Sep 07, 2005 at 05:17:13PM -0500, linas wrote:
> On Wed, Sep 07, 2005 at 06:14:41PM +0300, Andris Pavenis was heard to remark:
> >  
> > -TESTS = test
> > -
> >  check_PROGRAMS = test
> >  
> >  test_SOURCES = test.c
> 
> This is a really bad name for a unix prgram. For, on my machine
> 
>   which test 
> 
> returns
> 
>   /usr/bin/test

$ ./test

How else would you run a program called "test" in the current directory?

Cheers,

Phil

-- 
"Have you googled her yet?" "Willow, she's 17." "It's a search engine."
                                           Willow and Xander: Buffy: Help

Re: preparing a new release (part 2)

[linas]

On Wed, Sep 07, 2005 at 11:47:26PM +0100, Philip Kendall was heard to remark:
> On Wed, Sep 07, 2005 at 05:17:13PM -0500, linas wrote:
> > On Wed, Sep 07, 2005 at 06:14:41PM +0300, Andris Pavenis was heard to remark:
> > >  
> > > -TESTS = test
> > 
> > This is a really bad name for a unix prgram. For, on my machine
> > 
> >   which test 
> > 
> > returns
> > 
> >   /usr/bin/test
> 
> $ ./test
> 
> How else would you run a program called "test" in the current directory?

export PATH=$PATH:.

--linas

Re: preparing a new release (part 2)

[Philip Kendall]

On Wed, Sep 07, 2005 at 05:52:42PM -0500, linas wrote:
> On Wed, Sep 07, 2005 at 11:47:26PM +0100, Philip Kendall was heard to remark:
> 
> export PATH=$PATH:.

<rant>

This is a *very bad idea*. It leaves you open to typo trojans: say you're
in the /tmp directory on a multi-user box. You accidently type "sl"
instead of "ls". You now run /tmp/sl, which could have been installed by
anybody. You lose. "." should *never* appear on your $PATH.

</rant>

Phil (with vague apologies for off-topicness)

-- 
"I jump out of the circle, jump back in, and, and, shake my gourd."
"Hey, I think I know this ritual. The ancient shamans were next called
upon to do the Hokey-Pokey and to turn themselves around?"
                                            Giles and Buffy: Intervention

Re: preparing a new release (part 2)

[Stewart V. Wright]

[-- Attachment #1: Type: text/plain, Size: 830 bytes --]

G'day Linas,

* linas <linas@austin.ibm.com> [050907 17:18]:
> On Wed, Sep 07, 2005 at 06:14:41PM +0300, Andris Pavenis was heard to remark:
> >  
> > -TESTS = test
> > -
> >  check_PROGRAMS = test
> >  
> >  test_SOURCES = test.c
> 
> This is a really bad name for a unix prgram. For, on my machine
> 
>   which test 
> 
> returns
> 
>   /usr/bin/test


I would not disagree too loudly on aesthetic grounds, but automake is
smart enough to preface the check_PROGRAMS with a ./ (or at least
enough of the directory tree) so there is no real problem in calling
it test, true or even false (if you wanted to be _really_ nasty to the
next generation of developers).

All in all the naming is not really a worry.  I'm surprised about the
ordering of TESTS and check_PROGRAMS though...



Cheers,

S.

[-- Attachment #2: Type: application/pgp-signature, Size: 274 bytes --]

Re: preparing a new release (part 2)

[Brian Gough]

Andris Pavenis writes:
 > Builds OK for DJGPP (v2.04 beta, 2,03 not tested, gcc-4.0.1, WinXP Pro+SP2).
 > Also all tests passes if the following small EXEEXT related patch is applied 
 > (it was for 1.6, but it  applies also to 1.6.92):

Thanks, I will fix that.

-- 
Brian Gough

Re: preparing a new release (part 2)

[linas]

On Tue, Sep 06, 2005 at 12:22:01PM +0100, Brian Gough was heard to remark:
> I've made second version to fix some of the errors reported (memory leak in

It appears to pass configure; make; make check on 
Novel SuSE Enterpirse Linux SLES9 on a Power5 box.

--linas

Re: preparing a new release (part 2)

[Jeff Spirko]

GSL 1.6.92 passes all tests on IA-64 with gcc-3.3.3
under SuSE Linux Enterprise Server 9.

Regards,
-- 
Jeff Spirko   spirko@lehigh.edu   spirko@yahoo.com   WD3V   |=>

The study of non-linear physics is like the study of non-elephant biology.

All theoretical chemistry is really physics;
and all theoretical chemists know it. -- Richard P. Feynman 

Re: preparing a new release

[keith.briggs]

gsl-1.6.92.tar.gz passed all tests on AMD64 with gcc-4.0.1.

Keith

PS: perhaps the grammar in "All 1 tests passed" could be improved.

Re: preparing a new release

[Brian Gough]

keith.briggs@bt.com writes:
 > gsl-1.6.92.tar.gz passed all tests on AMD64 with gcc-4.0.1.
 > 
 > Keith
 > 
 > PS: perhaps the grammar in "All 1 tests passed" could be improved.

One for the Automake maintainers I think -- I've cc'ed this to
bug-automake@gnu.org.

The "make check" messages are from automake (maybe "All tests passed
(1/1)" would do the job).

-- 
Brian Gough